Project Summary/Abstract In studying the PI3 Kinase isoform dependence of different tumor types, we made an extremely surprising finding that now turns out to have considerable clinical importance. This finding forms the basis for this OIA application. We discovered that tumors driven by the loss of the PTEN tumor suppressor are uniquely dependent on the p110? isoform of PI3 Kinase. This finding likely explains why PI3K inhibitors first tested on PTEN null tumors failed in the clinic, as they were poor p110? inhibitors. New p110b specific compounds are now showing clinical promise. In attempting to understand the molecular mechanisms that uniquely couple PTEN loss to p110? activation, we have uncovered a set of molecular mechanisms, which not only explains how p110? is activated in response to PTEN loss but also suggests why the same tumors might quickly become partially or even totally resistant to PI3K inhibition. Notably the same mechanisms clearly suggest other drug targets, which can and should be attacked in combination with PI3K in PTEN null tumors. Our very recent data identify 2 proteins that uniquely interact with p110?, and not with p110?, form a positive feedback loop in the absence of PTEN. One of these proteins is the small ras family GTPase known as Rac, which interacts with p110? but not p110?. We have recently shown that Rac localizes p110b to the lipid rafts where it is activated. Thus Rac is an upstream activator of p110?. However Rac family members are unique in that their activators, the Rac GEFs, are activated by the phosphoinositide products of PI3Ks. Thus Rac is also a downstream effector of p110?. The interactions of Rac and p110? constitute the very definition of a positive feedback loop. However, this leaves open how the Rac/p110? feedback loop is initiated- what activates p110?/Rac in the first place. We have found that the activation event is dependent the small adapter protein CRKL which is also a p110? specific binding protein. Activation of CRKL occurs via a SRC/p130Cas signaling cascade that is also activated by PTEN loss. Notably SRC signaling renders cells resistant to PI3K inhibition. Finally and most exciting we have found that p110b inhibitors synergize with immune checkpoint blockade. We have generated data already showing the inhibitors of SRC RAC PAK (another downstream target of RAC) lipid raft formation and immune checkpoints can all combine well with p110b inhibitors in vitro, and in some cases, in vivo. This grant will focus more rigorous testing of new drug combinations on the one hand and on the other hand, further refining our mechanistic understanding of the effects of PTEN loss to generate even better combination therapy.